The Agent Benchmark

‹ Software

Top AI software engineer agents

Software and IT · 40 ranked

The top AI software engineer agents on The Agent Benchmark are Codex (9.7 of 10), Claude Code (9.7) and Cursor (9.7), of 40 ranked. Data as of 3 Oct 2026.

An AI software engineer writes or changes code in an existing codebase to build features, fix bugs, or migrate and modernize code. Examples include coding agents, AI coding assistants, and AI code editors. Not apps built from a prompt for non-programmers (AI app builder).

Ranked agents
40
in Software
Median score
2.5
of 10
From YC
32
8 from outside YC

Ranked software engineer agents

All 40, best score first. The score adds public proof, scale, momentum and autonomy; it is not a hands-on test. How scoring works

  1. 1Codexby OpenAIThe best way to build with agents. Codex accelerates real engineering work, from planning and building features to refactors, reviews, and releases.Proof10Scale10Momentum10Autonomy8Score9.7
  2. 2Claude Codeby AnthropicAnthropic's agentic coding tool for developers.Proof10Scale10Momentum10Autonomy8Score9.7
  3. 3Cursorby AnysphereCursor is your coding agent for building ambitious software.Proof10Scale10Momentum10Autonomy8Score9.7
  4. 4Factory Droidsby FactoryFactory Droids automate coding, testing, and deployment for startups and enterprises.Proof10Scale3Momentum10Autonomy8Score7.6
  5. 5Devinby CognitionCognition operates Devin, the first autonomous software engineer.Proof7Scale3Momentum10Autonomy8Score6.7
  6. 6Copilot cloud agentby GitHubGitHub Copilot works every step of the way, from an open issue to reviewing and merging a pull request.Proof0Scale10Momentum3Autonomy8Score5.0
  7. 7AlloyW23AI prototyping tool for product managersProof5Scale4Momentum5Autonomy6Score4.9
  8. 8OpenHandsMeet OpenHands, the open-source, model-agnostic platform for cloud coding agents.Proof2Scale2Momentum7Autonomy8Score4.2
  9. 9CosineW23Fully Agentic SWEProof0Scale4Momentum6Autonomy8Score3.9
  10. 10WozW25Claude Code plugin that reduces token consumption and cost by 50%Proof0Scale3Momentum7Autonomy6Score3.6
  11. 11MesmerSp25The operating system for the 100x Engineering TeamProof4Scale3Momentum2Autonomy6Score3.5
  12. 12TempoS23Visual Editor for React, powered by AIProof0Scale5Momentum4Autonomy6Score3.4
  13. 13HypercubicF25Agentic AI for Mainframe Operations and ModernizationProof2Scale2Momentum5Autonomy6Score3.4
  14. 14LightSprintSp26Software factory for product teamsProof0Scale2Momentum7Autonomy6Score3.3
  15. 15LaminarF24Enterprise systems delivery without the turbulence.Proof0Scale6Momentum2Autonomy6Score3.2
  16. 16Reason MachinesSp26Reason Machines builds Reason Agent, the autonomous software engineer…Proof0Scale1Momentum6Autonomy8Score3.0
  17. 17ReplicasSp26Run any coding agent harness in the cloudProof2Scale1Momentum3Autonomy8Score2.8
  18. 18OpenCodeW21The open source AI coding agentProof0Scale6Momentum0Autonomy6Score2.7
  19. 19AssembleS26Autonomous IT DeliveryProof0Scale2Momentum4Autonomy6Score2.5
  20. 20LinzumiSp26Team chat to command all your company's coding agentsProof0Scale2Momentum4Autonomy6Score2.5
  21. 21HopliteS26Effortlessly deploy cloud software factories.Proof0Scale1Momentum4Autonomy8Score2.5
  22. 22PartyS24Secure multiplayer software factory for your whole team in minutesProof2Scale3Momentum0Autonomy6Score2.4
  23. 23Adri AIW23Your AI Partner for SAP DevelopmentProof3Scale2Momentum0Autonomy6Score2.4
  24. 24ConductorS24Run a team of coding agents on your MacProof0Scale3Momentum2Autonomy6Score2.3
  25. 25CapyF24The IDE for the parallel ageProof0Scale3Momentum2Autonomy6Score2.3
  26. 26stagewiseS25The Open Source Agentic IDEProof2Scale1Momentum2Autonomy6Score2.3
  27. 27AmpAmp is a coding agent and development environment built for the frontier.Proof0Scale0Momentum4Autonomy8Score2.2
  28. 28JcodeS26High performance coding agentsProof0Scale0Momentum4Autonomy8Score2.2
  29. 29Human BehaviorSp25Understand your users beyond the numbersProof0Scale4Momentum0Autonomy6Score2.1
  30. 30SweepS23The best AI coding assistant for JetBrainsProof0Scale2Momentum2Autonomy6Score2.0
  31. 31OnlookW25Cursor for DesignersProof0Scale2Momentum2Autonomy6Score2.0
  32. 32ByteAskF26The AI coding agent for C and C++Proof0Scale1Momentum1Autonomy8Score1.8
  33. 33HumanLayerF24The multiplayer control plane for your software factoryProof0Scale1Momentum2Autonomy6Score1.7
  34. 34SparklesW26Make everyone on your team an engineerProof0Scale0Momentum3Autonomy6Score1.7
  35. 35nCompass TechnologiesW24Optimize performance on GPUs - 10x fasterProof0Scale2Momentum0Autonomy6Score1.5
  36. 36MagiCodeS24A Reliable AI Frontend EngineerProof0Scale2Momentum0Autonomy6Score1.5
  37. 37Double – Coding CopilotW23AI coding copilot engineered for performanceProof0Scale1Momentum0Autonomy6Score1.2
  38. 38NuancedW24the AI coding app for thinking.Proof0Scale1Momentum0Autonomy6Score1.2
  39. 39OpslaneS24Verify AI-generated code before you ship itProof0Scale0Momentum0Autonomy6Score0.9
  40. 40Random LabsS24Random labs is building long running autonomous coding agents.Proof0Scale0Momentum0Autonomy6Score0.9

Can agents do this job?

Most tasksin software engineering

Public benchmarks cover 47% of the work in this job family (9 benchmarks). About the job, not one agent.

Source: Can Agents Work?

Data as of 3 Oct 2026. The score uses public evidence only. Read the methodology · Missing an agent? Submit it

↑ ↓ to move · Enter to open